New debug kernel installation mechanism committed to HEAD

Greg 'groggy' Lehey grog at lemis.com
Sun Oct 24 08:11:30 PDT 2004


On Thursday, 21 October 2004 at  9:59:36 +0200, Richard Nyberg wrote:
> At Thu, 21 Oct 2004 01:59:13 -0400,
> Andrew Atrens wrote:
>>
>> Btw, is there a how-to somewhere on how one would load a kernel module
>> into gdb to do post-mortem on a vmcore? If one can't do this, then having
>> debug info in modules would definitely be less useful.
>
> Yes. See the FreeBSD developer handbook.
> http://www.freebsd.org/doc/en_US.ISO8859-1/books/developers-handbook/
> Section 11.7 in particular.
>
>> Should this work? -
>>
>> gdb -k kernel.0 vmcore.0
>> gdb> add-symbol-file /nvidia.ko 0xc060d000
>> gdb> bt
>
> I do this on a regular basis:
> gdb -k kernel.debug vmcore.9
>
> follow this chain to get the address of my module
> *linker_files.tqh_first.link.tqe_next.link.tqe_next
>
> objdump --headers /usr/arla/bin/nnpfs.ko | grep text
> to see the relocation address.
>
> add address and relocation address
>
> add-symbol-file /usr/arla/bin/nnpfs.ko sum-of-addresses

This is effectively what the 'asf' macro in my debugging macro
collection does.  But it's not always necessary.

Greg
--
Finger grog at xxxxxxxxx for PGP public key
See complete headers for address and phone numbers





More information about the Kernel mailing list